32 bit and 64 bit---- doubt
Hi all its a very basic ques may be for many of you .but i have this confusion from long that is. when we say this application is 32 bit application or 64 bit what do we mean by that is it processor which is 32 bit or 64 bit or the OS which is is 32 bit or 64 bit or what ? 2. Does it have some relation that in windows in C language int size is 2 bytes in linux gcc it is 4 bytes i mean any relation of word size with it ?? -- Thanks Regards Nidhi
Re: 32 bit and 64 bit---- doubt
On Mon, Sep 29, 2008 at 12:42 PM, nidhi mittal [EMAIL PROTECTED] wrote: Hi all its a very basic ques may be for many of you .but i have this confusion from long that is. when we say this application is 32 bit application or 64 bit what do we mean by that is it processor which is 32 bit or 64 bit or the OS which is is 32 bit or 64 bit or what ? It means the application can access 2^32 bits of memory i.e 4gigs.And this means a 64 bit application needs a 64 bit kernel atleast, which will run on a 64 bit processor. Aliter a 32 bit application can run on top of either 32 or 64 bit kernel. And a 32 bit kernel can run atop a 32 bit processor or 64 but processor. 2. Does it have some relation that in windows in C language int size is 2 bytes in linux gcc it is 4 bytes No such restriction for C in windows. It is still 4 bytes. i mean any relation of word size with it ?? word size is 32 bits on both Windows and Linux. Cu, --Pradeep -- Thanks Regards Nidhi -- play the game -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: About Bad Magic Number
On Tue, Sep 23, 2008 at 7:27 PM, mayank rana [EMAIL PROTECTED] wrote: Hi All, I am sending fasync from my driver to user space application. While sending fasync, I am getting below error : kill_fasync:bad magic number in fasync_struct ! Can anyone suggest any solution for this ? You fs superblock seems screwed up. Run fsck manually with a different back up superblock from the disk. Cu, --Pradeep Thanks. Regards, _-_Mayank Rana_-_ -- play the game -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: 32 bit and 64 bit---- doubt
Apart from these, following wiki link best describes what makes an architecture as 64 bit. http://en.wikipedia.org/wiki/64-bit On Mon, Sep 29, 2008 at 1:22 PM, pradeep singh [EMAIL PROTECTED]wrote: On Mon, Sep 29, 2008 at 12:42 PM, nidhi mittal [EMAIL PROTECTED] wrote: Hi all its a very basic ques may be for many of you .but i have this confusion from long that is. when we say this application is 32 bit application or 64 bit what do we mean by that is it processor which is 32 bit or 64 bit or the OS which is is 32 bit or 64 bit or what ? It means the application can access 2^32 bits of memory i.e 4gigs.And this means a 64 bit application needs a 64 bit kernel atleast, which will run on a 64 bit processor. Aliter a 32 bit application can run on top of either 32 or 64 bit kernel. And a 32 bit kernel can run atop a 32 bit processor or 64 but processor. 2. Does it have some relation that in windows in C language int size is 2 bytes in linux gcc it is 4 bytes No such restriction for C in windows. It is still 4 bytes. i mean any relation of word size with it ?? word size is 32 bits on both Windows and Linux. Cu, --Pradeep -- Thanks Regards Nidhi -- play the game -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: 32 bit and 64 bit---- doubt
On Mon, 2008-09-29 at 12:42 +0530, nidhi mittal wrote: Hi all its a very basic ques may be for many of you .but i have this confusion from long that is. when we say this application is 32 bit application or 64 bit what do we mean by that is it processor which is 32 bit or 64 bit Primarily it is that. or the OS which is is 32 bit or 64 bit Secondarily that too. But x86_64/amd64/em64t/intel64 has x86 compatibility (and Sparc64 for Sparc perhaps too). So can install and run a 64bit kernel and run the whole userspace compiled for 64bit. Or you can run only 32bit apps/libs in userspace on 64bit kernel (because your apps do not need 4GB address space but you have 64GB RAM in your hardware). or what ? Basically yes. But x86/64/AMD64/EM64T/... can run x86 32bit apps too (if the necessary libs are installed). 2. Does it have some relation that in windows in C language int size According to the C standard, sizeof(int) is the native register width. And that is basically defined by the CPU. So if sizeof(int) == 2, you probably have some 16bit microcontroller. And at least for WinNT and XP, sizeof(int) == 4 on normal PCs BTW. is 2 bytes in linux gcc it is 4 bytes i mean any relation of word size with it ?? Yup. Bernd -- Firmix Software GmbH http://www.firmix.at/ mobil: +43 664 4416156 fax: +43 1 7890849-55 Embedded Linux Development and Services -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: About Bad Magic Number
On Mon, 29 Sep 2008 13:24:28 +0530 pradeep singh [EMAIL PROTECTED] wrote: On Tue, Sep 23, 2008 at 7:27 PM, mayank rana [EMAIL PROTECTED] wrote: Hi All, I am sending fasync from my driver to user space application. While sending fasync, I am getting below error : kill_fasync:bad magic number in fasync_struct ! Can anyone suggest any solution for this ? You fs superblock seems screwed up. Run fsck manually with a different back up superblock from the disk. While it is true that filesystem superblocks contain magic numbers, it doesn't imply that every magic number has something to do with filesystem superblocks. Magic numbers are used throughout the kernel to check if structures passed to a function are indeed correct: if the magic number fails, it is a sign that something is wrong with the structure. It could be corrupted by something, it could also be that the user forgot to set the number. Without seeing the code, it's hard to tell what's wrong. Erik -- Erik Mouw -- [EMAIL PROTECTED] GPG key fingerprint: D6AC 7F15 A26E C5C4 62E0 4A58 FCF9 551C 9B48 B68D -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
network delay simulation
Hi all, Is there a mechanism to simulate random network latencies within kernel? for example do a net/scheduler or just FIFO delay within network driver? Regards, Bizhan
Re: 32 bit and 64 bit---- doubt
On Mon, Sep 29, 2008 at 2:38 PM, Bernd Petrovitsch [EMAIL PROTECTED] wrote: snip 2. Does it have some relation that in windows in C language int size According to the C standard, sizeof(int) is the native register width. And that is basically defined by the CPU. So if sizeof(int) == 2, you probably have some 16bit microcontroller. Or, you are running an x86 processor in virtual 8086 mode. For eg. running a DOS app under Linux or Windows. Kindly CMIIW. Best regards, Pranav http://pranavsbrain.peshwe.com
Re: 32 bit and 64 bit---- doubt
On Mon, 2008-09-29 at 17:38 +0530, Pranav Peshwe wrote: On Mon, Sep 29, 2008 at 2:38 PM, Bernd Petrovitsch [EMAIL PROTECTED] wrote: snip 2. Does it have some relation that in windows in C language int size According to the C standard, sizeof(int) is the native register width. And that is basically defined by the CPU. So if sizeof(int) == 2, you probably have some 16bit microcontroller. Or, you are running an x86 processor in virtual 8086 mode. For eg. running a DOS app under Linux or Windows. Kindly CMIIW. Is that actually possible? And with which toolchain? Bernd -- Firmix Software GmbH http://www.firmix.at/ mobil: +43 664 4416156 fax: +43 1 7890849-55 Embedded Linux Development and Services -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: 32 bit and 64 bit---- doubt
On Mon, Sep 29, 2008 at 5:46 PM, Bernd Petrovitsch [EMAIL PROTECTED] wrote: On Mon, 2008-09-29 at 17:38 +0530, Pranav Peshwe wrote: On Mon, Sep 29, 2008 at 2:38 PM, Bernd Petrovitsch [EMAIL PROTECTED] wrote: snip 2. Does it have some relation that in windows in C language int size According to the C standard, sizeof(int) is the native register width. And that is basically defined by the CPU. So if sizeof(int) == 2, you probably have some 16bit microcontroller. Or, you are running an x86 processor in virtual 8086 mode. For eg. running a DOS app under Linux or Windows. Kindly CMIIW. Is that actually possible? And with which toolchain? Bernd -- Firmix Software GmbH http://www.firmix.at/ mobil: +43 664 4416156 fax: +43 1 7890849-55 Embedded Linux Development and Services -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ Not sure if it does the same thing, Turbo C on windows shows sizeof(int) == 2. -- Sunil.
Re: 32 bit and 64 bit---- doubt
On Mon, Sep 29, 2008 at 5:46 PM, Bernd Petrovitsch [EMAIL PROTECTED] wrote: On Mon, 2008-09-29 at 17:38 +0530, Pranav Peshwe wrote: On Mon, Sep 29, 2008 at 2:38 PM, Bernd Petrovitsch [EMAIL PROTECTED] wrote: snip 2. Does it have some relation that in windows in C language int size According to the C standard, sizeof(int) is the native register width. And that is basically defined by the CPU. So if sizeof(int) == 2, you probably have some 16bit microcontroller. Or, you are running an x86 processor in virtual 8086 mode. For eg. running a DOS app under Linux or Windows. Kindly CMIIW. Is that actually possible? And with which toolchain? Yes, AFAIK. Windows supports running apps in VM86 mode out of the box. If Nidhi got sizeof(int) == 2, then it probably might have been done in Turbo C/C++ which runs in VM86 mode on windows. On linux, there are DOSEmu and DOSBox. Both worked well for me. I had played Doom 2 on DOSBox :) Best regards, Pranav http://pranavsbrain.peshwe.com
Re: network delay simulation
Bizhan Gholikhamseh (bgholikh) escreveu: Hi all, Is there a mechanism to simulate random network latencies within kernel? for example do a net/scheduleror just FIFO delay within network driver? Regards, Bizhan pktgen can help you. http://www.mjmwired.net/kernel/Documentation/networking/pktgen.txt Best regards Alan Menegotto -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: inode and blocks
Not an assignment actually, but a project. We are working on open hierarchical storage management, in which we store files on disks according to different file placement policies. For eg. if i say that all the important files, like all the employee database should be in disk 1 and all the songs on disk 2, then we place them accordingly in different disks. On Mon, Sep 29, 2008 at 2:31 AM, Theodore Tso [EMAIL PROTECTED] wrote: On Sun, Sep 28, 2008 at 03:49:04PM +0530, Rohit Sharma wrote: Suppose i have a file named abc.txt and i want to specify that all the *.txt files must be allocated between block groups no. 100 - 200 in ext2 fs. Is there a way to do this? can we modify function ext2_new_inode and find_group_orlov for this? You would have to modify kernel code to do this; the main question which comes to mind is *why* would you want to do something like this? It seems like an ideal problem set that a professor might give to a student, since it would force them to try to get from an inode to the pathname used to open the file. So it seems to be one of these really pointless things that isn't particularly useful in real life, except for pedagogical purposes. - Ted -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How to know the maximum memory pages that can be allocated?
On Mon, Sep 29, 2008 at 9:48 AM, Peter Teoh [EMAIL PROTECTED] wrote: I would like to know the maximum memory I can allocate (using alloc_page()) in GFP_KERNEL and __GFP_HIGHMEM memory area - how can I know that? Peter, I doubt you can precisely know that at any given time. :) A lot of issues come into picture, the order of pages you want to allocate, fragmentation, pages used as anonymous pages , pages for other dynamic allocation which is required by kernel etc. All these make it really difficult for precise answer. Approximation I think you do it using some simple maths. HTH, Thanks, --Pradeep Thanks. -- Regards, Peter Teoh -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ -- play the game -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: How to know the maximum memory pages that can be allocated?
thank you pradeep singh for the sharing. In the /proc/meminfo, I saw this (and my questions follows): MemTotal: 3046884 kB MemFree: 1502524 kB ===this is the physical free memory? Buffers: 90820 kB Cached: 624168 kB SwapCached: 0 kB Active: 964128 kB Inactive: 519928 kB HighTotal: 2162240 kB HighFree: 758656 kB===this is the total free memory available via using __get_free_pages(__GFP_HIGHMEM)? LowTotal: 884644 kB LowFree:743868 kB===this is the total free memory available via using __get_free_pages(GFP_KERNEL)? SwapTotal: 4192924 kB SwapFree: 4192924 kB Thanks for the sharing On Mon, Sep 29, 2008 at 11:50 PM, pradeep singh [EMAIL PROTECTED] wrote: On Mon, Sep 29, 2008 at 9:48 AM, Peter Teoh [EMAIL PROTECTED] wrote: I would like to know the maximum memory I can allocate (using alloc_page()) in GFP_KERNEL and __GFP_HIGHMEM memory area - how can I know that? Peter, I doubt you can precisely know that at any given time. :) A lot of issues come into picture, the order of pages you want to allocate, fragmentation, pages used as anonymous pages , pages for other dynamic allocation which is required by kernel etc. All these make it really difficult for precise answer. Approximation I think you do it using some simple maths. HTH, -- Regards, Peter Teoh -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
Re: why not new-style I2C eeprom driver?
Hello Srinivas, On Mon, Sep 29, 2008 at 10:33:31AM +0530, [EMAIL PROTECTED] wrote: Would like to know why the I2C eeprom driver is not upgraded to new-style? Have you checked the latest kernel sources? Since 2.6.27-rc there is AT24 which should meet your needs. Why is that some devices are still following legacy I2C driver style. Because nobody converted them yet. This is mainly a problem of not having the hardware for testing. Every converted driver is very welcome. If i want to write an I2C driver, is there any dependency on the device that makes me to write only legacy driver? Nope. Please always use the new style. Actually, i am trying to rewrite the I2C eeprom driver in new-style using linux kernel device model for my understanding. Then have a look at at24.c for references :) All the best, Wolfram Sang -- Dipl.-Ing. Wolfram Sang | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry signature.asc Description: Digital signature
Re: 32 bit and 64 bit---- doubt
32 bit. app 1) The addresses are limited to 32 bit(can be more in some cases). The address space is limited to this value. 2) It will run on both 32 bit OS and 64 bit OS(64 bit OS can run 32 bit apps) 3) A pointer is of 4 bytes. 4) some other differences wrt to other data types 64bit app: 1) the address are limited to 2^64 ( theoretically) 2) It will run only on 64 bit OS 3) A pointer is of 64 bits(8 bytes) BTW, Integer is always 4 bytes. 2 byte integers are history. Very few people use turbo C compiler nowadays. gcc is the way to go. -Sri On Mon, Sep 29, 2008 at 12:42 PM, nidhi mittal [EMAIL PROTECTED]wrote: Hi all its a very basic ques may be for many of you .but i have this confusion from long that is. when we say this application is 32 bit application or 64 bit what do we mean by that is it processor which is 32 bit or 64 bit or the OS which is is 32 bit or 64 bit or what ? 2. Does it have some relation that in windows in C language int size is 2 bytes in linux gcc it is 4 bytes i mean any relation of word size with it ?? -- Thanks Regards Nidhi -- SK Malik
Re: inode and blocks
On Mon, Sep 29, 2008 at 8:03 PM, Rohit Sharma [EMAIL PROTECTED] wrote: Not an assignment actually, but a project. We are working on open hierarchical storage management, in which we store files on disks according to different file placement policies. For eg. if i say that all the important files, like all the employee database should be in disk 1 and all the songs on disk 2, then we place them accordingly in different disks. On Mon, Sep 29, 2008 at 2:31 AM, Theodore Tso [EMAIL PROTECTED] wrote: On Sun, Sep 28, 2008 at 03:49:04PM +0530, Rohit Sharma wrote: Suppose i have a file named abc.txt and i want to specify that all the *.txt files must be allocated between block groups no. 100 - 200 in ext2 fs. Is there a way to do this? can we modify function ext2_new_inode and find_group_orlov for this? You would have to modify kernel code to do this; the main question which comes to mind is *why* would you want to do something like this? It seems like an ideal problem set that a professor might give to a student, since it would force them to try to get from an inode to the pathname used to open the file. So it seems to be one of these really pointless things that isn't particularly useful in real life, except for pedagogical purposes. - Ted -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ Hi Rohit, Just out of curiosity, how are you going to identify the type of file inside kernel ? from an extension or file format ? Thanks. -- Sunil.
.config versioning?
Hi there, I have 3 computers, they are all slightly different yet they all require pretty much the same basic options. However, I am still a newbie with the kernel and have found it tedious try to make such hardware work or have it work in such a way... to the point where I just forget what I did (as long as it works). See, I'm looking for a cleaner way!... I've had an idea by using a versioning system (hey, why not use git!) but on this single file, the .config. I was thinking on building a trunk which would include only the changes that are needed on the 3 computers and that have passed some testing phase (ie, that it really works well on the 3 of them). I would then have 3 branches, or sub-trunks that would complete the main trunk, but with specific changes for this computer. This way, for computer A, the final .config would = main trunk + A-trunk. This is all real fun git practice all in all, but I was wondering if there were a simpler way of dealing with this? If not, I'll go ahead and build myself a wrapper script to ease the use of git specifically for .config. Thanks, Simon -- To unsubscribe from this list: send an email with unsubscribe kernelnewbies to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ